ipcw_eif_update: Iterative IPCW Update Procedure of Efficient Influence...

Description Usage Arguments Details Value

View source: R/eifs.R

Description

Iterative IPCW Update Procedure of Efficient Influence Function

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
ipcw_eif_update(
  data_in,
  C,
  V,
  ipc_mech,
  ipc_weights,
  ipc_weights_norm,
  Qn_estim,
  Hn_estim,
  estimator = c("tmle", "onestep"),
  fluctuation = NULL,
  flucmod_tol = 100,
  eif_reg_type = c("hal", "glm")
)

Arguments

data_in

A data.table containing variables and observations of full data. That is, this corresponds to the data after application of a censoring process.

C

A numeric binary vector giving the censoring status of a given observation.

V

A data.table giving the values across all observations of all variables that play a role in the censoring mechanism.

ipc_mech

A numeric vector containing values that describe the censoring mechanism for all of the observations. Note that such values are estimated by regressing the censoring covariates V on the observed censoring C and thus correspond to predicted probabilities of being censored for each observation.

ipc_weights

A numeric vector of inverse probability of censoring weights. These are equivalent to C / ipc_mech in any initial run of this function. Updated values of this vector are provided as part of the output of this function, which may be used in subsequent calls that allow convergence to a more efficient estimate.

ipc_weights_norm

A numeric vector of the weights described in the previous argument. In this case, the weights are normalized.

Qn_estim

A data.table corresponding to the outcome regression. This is produced by invoking the internal function est_Q.

Hn_estim

A data.table corresponding to values produced in the computation of the auxiliary ("clever") covariate. This is produced easily by invoking the internal function est_Hn.

estimator

The type of estimator to be fit, either "tmle" for targeted maximum likelihood estimation or "onestep" for a one-step estimator.

fluctuation

A character giving the type of regression to be used in traversing the fluctuation submodel. The choices are "weighted" and "standard".

flucmod_tol

A numeric indicating the largest value to be tolerated in the fluctuation model for the targeted minimum loss estimator.

eif_reg_type

Whether a flexible nonparametric function ought to be used in the dimension-reduced nuisance regression of the targeting step for the censored data case. By default, the method used is a nonparametric regression based on the Highly Adaptive Lasso (from hal9001). Set this to "glm" to instead use a simple linear regression model. In this step, the efficient influence function (EIF) is regressed against covariates contributing to the censoring mechanism (i.e., EIF ~ V | C = 1).

Details

An adaptation of the IPCW-TMLE for iteratively constructing an efficient inverse probability of censoring weighted TML or one-step estimator. The efficient influence function of the parameter and updating the IPC weights in an iterative process, until a convergence criteria is satisfied.

Value

A list containing the estimated outcome mechanism, the fitted fluctuation model for TML updates, the updated inverse probability of censoring weights (IPCW), normalized versions of the same weights, the updated estimate of the efficient influence function, and the estimated IPCW component of the EIF.


txshift documentation built on Oct. 23, 2020, 8:27 p.m.